EP2166305B1 - Illumination unit and method for projecting an illumination pattern. - Google Patents

Description

The invention relates to a lighting unit with an optical pattern generating element arranged in the beam path of a light source and to a method for projecting a lighting pattern into a monitoring area according to the preamble of claims 1 and 10, respectively.

Cameras have long been used for surveillance and are increasingly being used in security technology. A typical safety application is the protection of a dangerous machine, such as a press or a robot, where protection is provided when a body part is engaged in a hazardous area around the machine. Depending on the situation, this can be the shutdown of the machine or the move to a safe position.

If the scene to be monitored is low-contrast or has areas with little structure, this makes the detection of safety-critical events more difficult or even impossible. This applies especially to three-dimensional monitoring procedures based on stereoscopy. Here, images of the scenery with a receiving system, which consists essentially of two spaced-apart cameras, won from slightly different perspectives. In the overlapping image areas, the same structures are identified and calculated from the disparity and the optical parameters of the camera system by means of triangulation distances and thus a three-dimensional image or a depth map.

Stereoscopic camera systems offer the advantage compared to conventional safety-related sensors such as scanners and light grids, nationwide depth information from a two-dimensionally recorded observation scenery determine. With the aid of the depth information, it is possible in safety-related applications to specify areas of protection that are more variable and more precise, and to distinguish more and more precise classes of permitted object movements. For example, it is possible to detect movements of the robot itself or passing movements of a body part on the dangerous machine in a different depth plane as non-hazardous. That would be indistinguishable from an impermissible intervention with a two-dimensional system.

In the context of safety technology, however, there is the additional requirement for a reliable safety function beyond the secure recognition of an intervention in the provided image data to produce this three-dimensional image data in the form of a dense depth map, ie a reliable distance value for each image area and preferably almost every pixel available to have. Large structureless areas or structure features similar to one another can prevent an unambiguous assignment of image areas when locating the correspondences between the structural elements of the images. At least two types of errors are conceivable, namely a failure to find corresponding structural elements or an incorrect assignment. The result is gaps in the three-dimensional images or erroneous calculations of the distances. Both are extremely dangerous for safety applications, as an inadmissible procedure could go undetected or an incorrect distance could be assigned. The shutdown of the source of danger may then be omitted, because the intervention is erroneously classified as uncritical.

The well localizable image features may either naturally be present for passive stereoscopy in the form of edges or textures in the scene under surveillance, or may be imprinted as an illumination pattern for active stereoscopy of the scene.

Passive stereo measurement systems have no influence on the quality and density of image characteristics. To a possible structural or contrast weakness occurs here simply lack of brightness as a possible source of error added. Therefore, in practice, depth information is only available selectively, which does not allow a reliable safety function. Of course, for a given application it is possible to select or prepare the scenery accordingly, but this is cumbersome and creates additional dependencies. Passive systems are not active in every respect At least in the ideal case, the minimum object size is limited only by the resolution of the stereo camera, and transitions between upstream object and background are recognized as good features for the stereo correlation. Such images also correspond to the natural visual impression and are suitable for visualization for diagnosis or configuration.

Active stereo measurement systems use, in addition to the receiving system, a lighting unit whose function is to illuminate the observed scene with structured light and thus to generate features through the numerous light-dark transitions in the images, by means of which the stereo algorithm reliably and comprehensively distance data from the scenery can extract. Such patterns may also be self-dissimilar, that is to say that for the present monitors there are in the first place no regions which are symmetrical relative to displacements. With the help of optical filters, the pattern can be made more visible and at the same time an override of the image sensors can be avoided.

The illumination pattern enables largely comprehensive depth information. The disadvantage is that a visual orientation is almost impossible because of the projected pattern and such an image is therefore less suitable as a basis for the visualization. In addition, difficulties arise in some situations to accurately detect object edges, such as occlusions at transitions from upstream objects in the background. In order to maintain security, a minimum object size with security buffer must be selected so that the effective resolution decreases.

Conventionally, lighting units with projection techniques similar to a slide projector or based on diffractive optical elements (DOE) are proposed as pattern generators. Projectors have the disadvantage that the pattern generation is in principle a selection process of the emitted light, in which an intermediate image is imaged in the form of a slide via a lens in the outer or observation room. As a result, the light output is very poor at about 30%. This is particularly problematic because the preferred because of their narrow band and longevity compared to halogen lamps used semiconductor light sources anyway hardly reach the required optical output power for more than shortest ranges. In addition, the pattern in the outer space only becomes sharp in one plane and blurs in front of the light intensity of the imaging lens more or less fast. Since you do not know the distances in dynamic sceneries in advance, it is not possible to define a single, sharp image plane. Finally, the effort in terms of the optics, such as the imaging lenses, the condenser lenses, homogenizations and the like, as well as in terms of electronics, such as the control of the light source or cooling, very high and a projector solution is therefore quite costly.

Although lighting units based on DOEs have significantly higher luminous efficiencies. However, they require well-collimated laser light and always transmit a portion of the order of 1% of the incident light unbent in the zeroth order. Thus, a still very concentrated beam leaves the illumination unit, and no solutions are known for removing this beam from the illumination field without further ado. Thus, such lighting units are either not able to meet requirements for eye safety or the laser protection class, and thus inadmissible for operation, or you have to make do with very little useful light, so that the unbent transmitted beam for eye safety remains uncritical. However, such low-power light does not allow generation of dense depth maps with stereoscopic algorithms over more than shortest ranges.

In the US 2007-0263903 A1 A stereoscopic camera system generates a structured illumination pattern by means of a lighting unit, which is then used to calculate distances. The pattern is created by illuminating a diffractive optical element with a laser. The problem of eye safety discussed above is not addressed here, as well as safety-related applications. However, conventional lightings of this type have a relatively low optical output power of less than one watt, and therefore the transmitted beam in the zeroth diffraction order is less critical. The resulting brightness is also generally sufficient in non-safety applications because occasional errors in the range calculation or portions for which no distance data can be temporarily calculated do not have serious consequences. Incidentally, there is always the possibility to increase the exposure times and thus to compensate for short ranges of active lighting. This is not possible in safety engineering, because in every very short Response period, several complete depth maps must be available in order to guarantee reliable evaluation with comparisons and multiple backups.

In the DE 10 2006 001 634 B3 A method for the stereoscopic creation of a distance image is described in which the scene is recorded twice, namely once with the imprint of a random pattern and a second time under a changed illumination. The stereo algorithm is then based on images in which the pixels do not contain the absolute values, but brightness quotients of the two images. A number of illuminations are described to generate the random pattern, such as masks, irradiated LCD displays or grids. None of the illuminations allows a sufficiently intense self-similar illumination pattern, the structure of which can be changed in a simple manner.

From the DE 10 2007 036 129 B3 It is known that each evaluated distance image is based on two three-dimensional images which have been produced under two different recording conditions, for example under illumination with a self-similar or irregular pattern and with the illumination switched off. The lighting device itself is conventionally based on a DOE and therefore has the above-mentioned disadvantages. It is also unable to vary the structure of the illumination pattern in a simpler manner than, for example, replacing the DOE.

The WO01 / 051886 A1 describes a surface contour measurement in which three-dimensional position information about surface points is determined by means of an image sensor. For this purpose, the surface is illuminated with an interference pattern resulting from the superimposition of two light source points. Several embodiments are presented, how two light source points can be generated from a physical light source, wherein in one embodiment the two main maxima of a diffraction grating are directed through a lens and an optional double slit.

From the US 2008/0106746 A1 For example, a method for three-dimensional gesture detection is known. For this purpose, an uncorrelated spot pattern is projected onto an object, for example the hand of a computer user. The pattern is created by dropping laser light onto a DOE or optical diffuser element which causes phase changes of the transmitted light. Elements of this pattern are in a reference pattern searched for depth information, the reference pattern being obtained in advance at a level of known distance. In other embodiments, the pattern is depth-dependent, so that the pattern can already be deduced from the structure of the pattern.

The DE 20 2007 010 884 U1 describes a monitoring device with tracking of the image sensor to keep a hazardous moving machine part in the field of view. Numerous two- and three-dimensional imaging methods are disclosed. In some embodiments, a regular or self-dissimilar illumination pattern is irradiated onto a robotic arm to be monitored.

In the US 4,742,237 a surface analysis by means of a moving grating is disclosed. The striped pattern of a grid is projected onto the surface by means of a lens. The lens may be formed as a zoom lens to vary the distances between the light and dark lines of the illumination pattern.

It is therefore an object of the invention to provide an improved lighting unit for the generation of a pattern structured.

This object is achieved by a lighting unit according to claim 1 and a method for projecting a lighting pattern into a monitoring area by means of a pattern generating element according to claim 10. The solution according to the invention is based on the basic idea of picking up a light pattern in the near field of the pattern generation element and projecting it into the monitoring space with an imaging objective. If the projection lenses have a short focal length, then the fact that the image sharp images the pattern only in a certain depth of field range is not serious.

Since the pattern in the near field changes very rapidly with the distance, even small changes in the focal plane are sufficient to obtain a very different pattern. This also makes it possible to switch very quickly and easily between different illuminations and thus to combine the advantages of the two complementary stereo methods, namely active and passive stereoscopy. This achieves a more stable detection method and a better object resolution. The need for an additional camera for diagnosis and configuration is eliminated, since the system itself supports the visualization of scene information. The lighting unit is flexibly controllable and can generate flexible patterns. It is very compact, inexpensive and durable.

As adjusting any electrical or mechanical adjustment options come into consideration, which can operate both continuously and discretely. Although preference is given to changing the setting during operation, the invention also encompasses performance forms in which the focal plane is already fixed at the desired distance in the course of production, for example by housing elements, in order to later produce the intended pattern during operation. The position of the focal plane does not necessarily have to be caused by physical changes, that is to say the effective focal plane, which, for example, is also shifted by light of a different wavelength.

Advantageously, a switching element is provided in order to switch over at least between a first illumination and a second illumination of the monitoring area, which differs from the first illumination, by changing the position of the focal plane of the imaging objective. The invention is not limited to two lights, third and other lights are adjustable by each other position of the focal plane. In connection with a camera, it is conceivable that each evaluated individual image is based on one recording under the different illuminations, but also a change of lighting only occasionally for testing purposes, or even only on special user input to the diagnosis or configuration.

Suitable pattern generation elements are those which generate a sufficiently variable distance-dependent intensity distribution in their beam path. This is not the case with a conventional mask or a slide, and a shift in the focal plane will only result in blurring, which will result in even disadvantageous contrast losses, but will not cause a structural change in the illumination pattern. Instead, according to the invention, the coherence of the light source is exploited as pattern-generating. There is only a redistribution of energy, no suppression, so that a very high efficiency can be achieved with a luminous efficacy of 90%.

In this case, the pattern generating element according to the invention an optical phase element with unevenness, which are designed to local phase differences between adjacent to the phase element incident light components and thus to generate the illumination pattern by interference. Such a phase element, also referred to as a phase plate or phase structure, is based on the phase contrast method in microscopy. A surface relief structure is applied to at least one surface of the phase element. A transmitted wavefront undergoes a phase difference between lands and trenches that results in localized intensity modulation when illuminated with coherent light. This brightness distribution is projected from the near field into the image plane, ie the monitoring area.

A conceivable alternative to the phase element is a diffractive optical element, which is then not used by the Nahfeldabbildung according to its purpose. A DOE carries a microstructure precalculated for a particular pattern, for example based on a Fourier transform, with which the incident collimated laser light beam is "copied" by diffraction effects into secondary orders. In the far field a deterministic pattern is created, for a wavelength and a DOE always the same pattern. According to the invention, the DOE is as it were misappropriated, namely not necessarily illuminated with collimated laser light, and above all, the near field is imaged, so before the special pattern of the DOE can form at all. Only by this mapping of the near field, the desired effect according to the invention results that the pattern can be changed.

Phase element and DOE are thus similar in that, due to unevenness of a phase or microstructure when illuminated with coherent light in the near field forms a variable with the distance pattern. The underlying effect, however, is once interference in a DOE diffraction and a phase element, and feature sizes are different.

The unevennesses are advantageously provided over at least the entire area of the pattern-generating element illuminated by the light source and / or the pattern-generating element does not have any smooth subregions. This creates phase differences or diffractions and thus interference patterns throughout the relevant surveillance area. Gaps or low-contrast areas are prevented and thus the basis for a nationwide surveillance or a dense depth map is laid.

The unevennesses are preferably irregular and very small in relation to the geometrical extent of the pattern-generating element, in particular its thickness, the unevenness varying over the pattern-generating element and / or the unevenness being provided on one or more surfaces of the pattern-generating element. The unevenness only has to produce phase differences in the range of the wavelengths of the light used, for which very small differences in thickness are sufficient. The bumps should preferably vary over the pattern generating element to prevent local as well as global regularities of the generated pattern. It is sufficient in principle to produce the phase differences already on a surface or surface of the pattern-generating element, but it can also have more surfaces unevenness, which of course are only optically active when they are in the light incidence area. Even production-related unevenness can be satisfied by omitting a smoothing step of the optically active surfaces. Then, the pattern generating element is particularly inexpensive. Likewise, the bumps can also be actively applied.

The pattern-generating element advantageously has transmissive or reflective properties, in particular a phase plate or a phase mirror. Depending on the desired beam path and spatial conditions, for example within a housing, a mirror or a plate is more suitable.

The pattern-generating element preferably comprises glass or plastic, is in particular a glass plate, a plastic casting or a foil, wherein the pattern-generating element is again preferably coated on at least one surface or has a foil, wherein the coating, foil or their connection region with the pattern-generating element, the local phase differences Has diffraction-generating bumps. This coating or the film may have a microstructure. As a glass plate or film, an extremely inexpensive pattern generating element can be produced. The required unevenness can be easily produced in these materials and if necessary additionally or alternatively as a film.

The adjusting element according to the invention is adapted to adjust the position of the focal plane so that the near field of the pattern generating element is projected into the monitoring area and generates a self-similar pattern there. This pattern is at least local, if not globally self-dissimilar, that is pattern structures Repeat at the earliest at a certain distance or outside an environment. In the case of using the illumination unit with a stereoscopic sensor, this environment is predetermined by the environment within which the stereo algorithm searches for correspondences in the two images. The focal plane is typically set to 10-500 microns or even a few millimeters to centimeters behind the microstructure. The choice of the focal plane influences the contrast and "rounding" of the structural elements, so that these properties can be adjusted via the position of the focal plane. Within limits, the position of the focal plane can also influence the structure size of the illumination pattern in the monitoring area, wherein the feature size is advantageously chosen to be at least so small that a pattern element is at most as large as the desired resolution of the sensor.

The adjusting element is preferably designed to lay the focal plane in the pattern-generating element itself, in particular in the plane of a microstructure of the pattern-generating element, and thus to produce a homogeneous intensity distribution in the monitoring region. This makes it easy to switch to a uniform illumination for the recognition of natural structures and visualization. Instead of switching between self-similar pattern and homogeneous illumination, which combines passive stereoscopy with active stereoscopy, changes between two differently structured patterns are conceivable in order to create a different basis for stereoscopic image data and thus to diversify the process.

The adjustment element preferably has a focusing device of the imaging objective and / or a distance adjustment device between the imaging objective and the pattern generation element, in particular for the relative displacement of the imaging objective relative to the pattern generation element or for displacing a further optical element between the pattern generation element and the imaging objective. In addition to a focus adjustment of the imaging lens itself, For example, by a piezo drive, the focal plane relative to the near field are also physically changed by changes in distance. But it is also conceivable to change only the optically effective path, for example by inserting a wedge plate, so that the light passes through a change in the setting of the focal plane a thicker or thinner area with a different refractive index. It is important that the change in the situation is fast and reproducible. This can also be achieved by using light sources of different wavelengths when switching over, because the focal plane of the imaging objective is wavelength-dependent. The disadvantage of this is that the lighting can then no longer be selected with a narrow-band filter because the filter must be permeable at least for the two selected light sources.

The light source preferably has at least one laser, wherein distribution optics, which can widen the laser light for illuminating the pattern-generating element, or collecting optics is provided between the laser and the pattern-generating element in order to focus the laser light onto the pattern-generating element. Coherent and divergent light, as it arises in a distribution optics, avoids possible problems with eye protection from the beginning. Because of the low power losses in the pattern generating element, moderate light outputs are sufficient or higher ranges can be achieved with high-power lasers than conventional ones. The light source is narrow band, so the useful light can be easily filtered out of the ambient light.

In one development of the invention, a multiplicity of illumination units are provided for illuminating a larger illumination area or for generating a larger illumination intensity, wherein the illumination areas of the individual illumination units add together in an overlapping or non-overlapping manner. The lighting units can tile-like produce a larger lighting area in the one alternative. The arrangement has the advantage that the light energies deposited in the monitoring area add up, without weakening pattern edges due to overlaps. As with multiple illumination units with overlapping illumination fields, multi-emitter laser diodes, which are commonly used for high output powers, are not an adequate means because they overlap multiple patterns and thus weaken the contrast. Since it is therefore not possible to arbitrarily increase the light power falling on the entire surveillance area, in this embodiment the power density is provided by smaller ones Increase lighting fields and achieve the desired size of the surveillance area through multiple lighting modules. It is conceivable to use in each case the same pattern generating element of all or more light sources, if it is large enough so that said disturbing effects of the light sources are mutually sufficiently suppressed. In the other alternative, identical or non-random patterns are superimposed, so that the required illumination intensity is achieved only in the common superimposition in order to meet the requirements of the eye protection with each individual illumination source and thus also overall.

The illumination device according to the invention is preferably used in a stereoscopic 3D camera, wherein an evaluation unit is provided, which is designed for the application of a stereo algorithm in which to generate a three-dimensional distance image mutually associated partial areas of the captured by the two cameras of the stereo camera images and their Distance is calculated on the basis of disparity. Due to the high efficiency of the lighting unit, high ranges are possible. For the dense depth maps made possible in accordance with the invention, it is not important whether the illumination pattern is known in advance, since it is only important to find structures of the same illumination pattern in images of the individual cameras taken at the same time in order to detect correspondences.

In this case, the evaluation unit is preferably designed to generate a three-dimensional distance image under the first illumination and the second illumination and to compare or compute the resulting three-dimensional image data with one another. The advantages of active and passive stereoscopy can be combined as described, and it is possible to create a diverse system.

The 3D camera is preferably designed as a security camera, wherein the evaluation unit can detect impermissible interventions in the surveillance area and then generate a shutdown signal, and wherein a safety output is provided, via which the security camera can output a shutdown signal to a monitored machine. An impermissible intervention means that, on the one hand, an intervention is detected and, on the other hand, this intervention does not correspond to a previously learned or algorithmically recognized permitted object, no permitted movement or the like. In this case, it is conceivable to precede the shutdown with a warning, for example to define protective fields in which a shutdown should take place, and upstream warning fields in which initially an example visual or audible alarm is triggered by the intervention leading to the shutdown possibly due to the warning to prevent. Security technology is particularly demanding, and a security camera can do its job better if sufficient contrast in the surveillance area is guaranteed and different lighting is available.

The method according to the invention can be developed in a similar manner and shows similar advantages. Such advantageous features are described by way of example but not exhaustively in the subclaims following the independent claims.

Fig. 1

eine schematische räumliche Gesamtdarstellung einer Ausführungsform einer erfindungsgemäßen Beleuchtungseinheit in einer 3D-Sicherheits- kamera;

Fig. 2

eine schematische Darstellung einer Ausführungsform der erfindungsge- mäßen Beleuchtungseinheit mit Abbildung des Nahfelds zur Erzeugung ei- nes strukturierten Beleuchtungsmusters;

Fig. 3

eine Darstellung gemäß Figur 2 mit Abbildung der Ebene eines Musterer- zeugungselements zur Erzeugung einer homogenen Beleuchtung;

Fig. 4

eine Darstellung gemäß Figur 2 zur Veranschaulichung einer Abstandsän- derung zwischen Mustererzeugungselement und Abbildungsobjektiv zur Änderung der Lage der Brennebene;

Fig. 5

eine Darstellung gemäß Figur 4 mit einer alternativen Ausführungsform der Erfindung, bei der zur Änderung der Lage der Brennebene eine Keilplatte eingeschoben wird; und

Fig. 6

eine Darstellung einer Anordnung mehrerer erfindungsgemäßer Beleuch- tungseinheiten zur Ausleuchtung eines größeren Überwachungsbereichs.

The invention will be explained in more detail below with regard to further features and advantages by way of example with reference to embodiments and with reference to the accompanying drawings. The illustrations of the drawing show in:

Fig. 1

a schematic overall view of an embodiment of a lighting unit according to the invention in a 3D security camera;

Fig. 2

a schematic representation of an embodiment of the inventive illumination unit with imaging of the near field to produce a structured illumination pattern;

Fig. 3

a representation according to FIG. 2 mapping the plane of a pattern-generating element to produce a homogeneous illumination;

Fig. 4

a representation according to FIG. 2 to illustrate a change in distance between pattern generating element and imaging lens to change the position of the focal plane;

Fig. 5

a representation according to FIG. 4 with an alternative embodiment of the invention, in which a wedge plate is inserted to change the position of the focal plane; and

Fig. 6

a representation of an arrangement of several inventive lighting units for illuminating a larger surveillance area.

FIG. 1 shows in a schematic three-dimensional representation of the general structure of a 3D security camera according to the invention 10 according to the stereoscopic principle, which is used for safety monitoring of a room area 12. Two camera modules are mounted in a known fixed distance from each other and each take pictures of the space area 12. In each camera, an image sensor 14a, 14b is provided, usually a matrix-shaped recording chip, which receives a rectangular pixel image, for example a CCD or a CMOS sensor. Associated with the image sensors 14a, 14b is an objective with imaging optics, which are shown as lenses 16a, 16b and can be realized in practice as any known imaging objective. The viewing angle of these optics is in FIG. 1 represented by dashed lines, each forming a viewing pyramid 18a, 18b.

In the middle between the two image sensors 14a, 14b, a lighting unit 100 is shown, this spatial arrangement is to be understood only as an example and the lighting unit may also be arranged asymmetrically or even outside the security camera 3D. The illumination unit 100 according to the invention generates a structured illumination pattern 20 or homogeneous illumination in an illumination area 102 in the spatial area 12 and will be described below in various embodiments in connection with FIGS FIGS. 2 to 6 explained in more detail.

A controller 22 is connected to the two image sensors 14a, 14b and the illumination unit 100. By means of the controller 22, the structured illumination pattern 20 is generated and varied in its structure or intensity as needed, and the controller 22 receives image data from the image sensors 14a, 14b. From this image data, the controller 22 computes three-dimensional image data (range image, depth map) of the spatial region 12 by means of a stereoscopic disparity estimation. The structured illumination pattern 20 thereby provides good contrast and a clearly assignable structure of each pixel in the illuminated spatial region 12. It is self-similar; the most important aspect of the self-dissimilarity being the at least local, better global lack of translational symmetries, so that no apparent shifts of picture elements in the different perspective images are detected due to the same illumination pattern elements that would cause errors in the disparity estimation. Alternatively, the controller may set the illumination unit 100 to provide a homogeneous illumination in which natural structures for the stereo algorithm and, for diagnostic or configuration purposes, also for the human eye are more visible.

With two image sensors 14a, 14b there is a known problem that structures along the epipolar line can not be used for the disparity estimation, because here no triangulation angle occurs or, in other words, the system can not locally distinguish whether the structure in the two images due to the perspective shifted against each other or whether only an indistinguishable other part of the same, aligned parallel to the base of the stereo system structure is compared. To solve this, in other embodiments, one or more other camera modules can be used, which are arranged offset from the connecting line of the original two camera modules.

The space area 12 monitored by the security sensor 10 may contain known and unexpected objects. This may, for example, be a robot arm, a machine, an operator and others. The room area 12 provides access to a source of danger, be it because it is an access area or because a dangerous machine is in the room area 12 itself. To safeguard this source of danger, one or more virtual protection and warning fields can be configured. Due to the three-dimensional evaluation, it is also possible to define these fields three-dimensionally, so that a great deal of flexibility arises. The controller 22 evaluates the three-dimensional image data for impermissible interventions. For example, the evaluation rules may stipulate that no object may exist in protected fields at all. More flexible evaluation rules provide differentiation between allowed and non-permitted objects, such as trajectories, patterns or contours, speeds or general workflows that are both learned in advance as well as evaluated on the basis of assessments, heuristics or classifications during operation.

If the controller 22 detects an impermissible intervention in a protective field, a warning or shutdown device 24, which in turn can be integrated into the controller 22, issues a warning or safeguards the source of danger, for example a robot arm or another machine. Safety-relevant signals, that is, in particular the switch-off signal, are output via a safety output 26 (OSSD, Output Signal Switching Device). It depends on the application, whether a warning is sufficient, or it is a two-stage protection provided in the first warned and is turned off only with continued object engagement or even deeper penetration. Instead of a shutdown, the appropriate Reaction also be immediate spending in a safe parking position.

To be suitable for safety applications, the sensor 10 is designed fail-safe. This means, inter alia, that the sensor 10 can test itself in cycles below the required response time, in particular also detects defects of the illumination unit 100 and thus ensures that the illumination pattern 20 is available in an expected minimum intensity, and that the safety output 26 and the Warn - or shutdown 24 safe, for example, two channels are designed. Likewise, the controller 22 is self-confident, thus evaluating two-channel or uses algorithms that can check themselves. Such regulations are standardized for general non-contact protective devices in EN 61496-1 or IEC 61496. A corresponding standard for security cameras is in preparation.

The security camera 10 is surrounded by a housing 28 and protected. Through a windscreen 30, light can pass in and out of the space area 12. The windshield 30 has filter characteristics which are matched to the transmission frequency of the illumination unit 100. Namely, even with high-intensity illumination units 100, under adverse conditions, such as very bright environments, large surveillance volumes or large distances of 5 meters and more, the useful signal in the noise of the entire spectrum can not be sufficiently detected. With a lighting unit 100 that emits only light in one or more narrow visible, ultraviolet, or infrared bands, and a matched filter 30, the signal-to-noise ratio can be significantly improved because ambient light outside of those bands is no longer relevant. The optical filter 30 can also be realized differently, for example in the lenses of the camera modules.

FIG. 2 shows a first embodiment of the lighting unit 100 according to the invention based on a near-field image of a complex interference pattern. As throughout the description, like reference numerals designate the same features. A light source 104 generates coherent light, which falls on an optic 106, which is exemplified here as a converging lens. This creates a limited coherent light beam. Alternatively, a distribution optics is conceivable, for example, with a scattering lens, then form light source 104 and optics 106 together a source of divergent coherent light, or you can also dispense with the optics 106, if the geometry allows or resulting light losses are acceptable.

The light is incident on a pattern-generating element 108, in the illustrated embodiment a phase element or a phase plate. It may be, for example, a simple thin glass plate whose one or both surfaces have a slightly uneven topology. The local thickness differences of the glass impose local phase shifts on the plane or weakly curved wavefront of the incident light. The dimensions of these local phase patterns are of the order of magnitude of the wavelength of light. After the pattern generating element 108, the light field then propagates without further influencing and gradually forms an intensity pattern with bright areas of constructive and dark areas of destructive interference. In the vicinity of a few millimeters behind the pattern generating element 108, the pattern changes from initially completely homogeneous intensity distribution over an intensity distribution similar to the surface structure of the pattern generating element 108 to a far field distribution which is highly structured and very stable with further increasing distance.

The interference pattern is strongly modulated with good spatial coherence of the light source 104 and appropriate surface structure of the pattern generating element 108, so that a nationwide interference pattern of highly visible and self-similar light-dark transitions. The good coherence leads to strong contrasts of the pattern and only a little homogeneous background light. With poor coherence or overlapping independent patterns, for example because of multiple light source points, the contrast is significantly reduced. The stereo algorithm may be affected and the security camera 10 is less robust to extraneous light.

An imaging lens 110, which can be configured in any known manner, for example as a simple converging lens or as a slide projector, forms with a focal plane 112 the near field with the interference pattern behind the pattern generation element in the monitoring area 12, so that the self-similar illumination pattern 20 is projected there , The interference pattern first forms in the near field immediately behind the pattern generation element 108, ie it varies with the distance to the pattern generation element 108. The position of the focal plane 112 therefore influences the structure of the resulting illumination pattern 20.

In addition to the imaging lens 110, an element for changing the natural illumination field of the illumination device 100 and the light source 104 may be provided be, for example, a cylindrical lens, or a diaphragm for limiting the illumination field. Such elements can be adjusted to achieve the desired illumination field.

Instead of a phase element, a DOE can also be used as pattern generation element 108. This DOE is then misused, as already explained, since the near field is imaged before the dedicated pattern of the DOE can form in the far field. Furthermore, the lighting is not as usual with a highly concentrated light beam, because the light of the light source 104 is expanded in the optics 106 to a divergent light beam or at least over a larger area. There are no problems with the eye protection.

The coherent light source 104 is preferably a laser diode, more preferably a single-emitter laser, because multiple source points form further interference sources and thus could cause contrast losses in the pattern. Because of the high light output in the phase element 108, comparatively moderate output powers are sufficient, but the strongest available high-power lasers can be used, especially for longer ranges.

The pattern generating element 108 shows unevenness at least over those areas in which the structured illumination pattern 20 is to arise. Smooth areas prevent interference and thus provide pattern-free areas that are undesirable, at least in most applications, because they can lead to gaps in the depth maps. A simple realization is that the unevennesses are formed on one side or on both sides over the entire surface of the phase element 108, or at least over the optically active surface. The unevenness should continue to be irregular, because otherwise interference structures with one another could arise in the illumination pattern 20.

The phase element 108 may be made of glass other than a transparent material, such as plastic or be formed as a film. The bumps can already be just the manufacturing tolerances of non-smoothed surfaces, or they are specially applied. Coatings or coating films can be applied so that just in ordinary optical applications considered insufficiently accurate uneven connections for the necessary unevenness provides. A coating with corresponding unevenness is conceivable, even in the form of microstructures. Especially the latter is a rather high overhead, and the phase element 108 shows the desired effects even with much more easily caused bumps. Plastic parts with any desired unevenness can be produced for example by injection molding.

The pattern properties can be specified more precisely by precisely specifying the surface structure of the phase element. For example, defined structures can be determined in simulation calculations and then generated by means of laser lithography. The structure thus produced can be replicated by embossing, molding with optical polymers or by precision injection molding.

The fill factor of the illuminated area of the phase element 108, that is to say the proportion of structural elements, should be approximately 50% or more for good contrast and efficiency. The structural elements themselves are preferably anisotropic in order to limit isotropies in the illumination pattern 20 and thus the risk of ambiguity. The structural elements may also be formed so that the optical power is distributed in a predetermined or optimal manner over the illumination area 102, for example, a Randabununkelung about given by the geometries or compensate for lens influences.

The focal plane of the imaging lens 110 is adjustable. In the embodiment according to FIG. 2 For this purpose, the control is connected to an electrical focus adjustment of the imaging objective 110. This makes it possible to change the structure and contrast of the illumination pattern 20 by varying the position of the focal plane 112. The location of the focal plane 112 is not to scale; in practice, the self-similar pattern may be tapped directly behind the pattern generator 108.

FIG. 3 FIG. 12 shows a situation where the focal plane 112 is placed in the plane of the microstructure of the pattern generating element 108. At this point, no interference is still effective, and therefore in the space area 12 creates a homogeneous illumination 21. It is of course also a focal plane 112 still further forward, so be selected within the pattern generating element 108 or in front of the pattern generating element 108. The switching between self-similar pattern and Therefore, the controller 22 achieves homogeneous illumination solely by shifting the focal plane.

This makes a particularly uncomplicated and fast switching between different lighting or lighting scenarios possible. This can be exploited in a variety of ways, for example, to set a desired feature size for the illumination pattern to use for diversification several selbstunähnliche pattern or to combine passive and active stereoscopy by switching between homogeneous illumination and selbstunähnlichem illumination pattern advantageous.

In FIG. 4 an additional adjustment element 114 is provided to change the relative distance between pattern generating element 108 and imaging lens 110. The adjusting element 114 can be realized for example via piezoelectric crystals. Although this does not change the focus setting of the imaging objective 110, it does, however, change the relative position of the focal plane 112 relative to the pattern generation element 108 FIGS. 2 and 3 can be changed over the adjusting element 14 between a position with homogeneous illumination 21 and different illumination patterns 20.

Another way to change the relative position of the focal plane 112 is in FIG. 5 shown. An adjustment member 116 advances an optical element, here a transparent wedge plate 118, further into the optical path between pattern-generating element 108 and imaging lens 110 or further out of the optical path. Thus, the light path is lengthened or shortened, which in turn effectively results in the near field being imaged from a different distance from the pattern generation element 108. Instead of a wedge plate, every optical element that changes the length of the light path depending on its position in the optical path is suitable.

Further possibilities are conceivable for changing the relative position of the effective focal plane, ie for imaging a differently spaced plane of the near field. In addition to any combination of the embodiments described, it is also conceivable, for example, to switch back and forth between light sources of different wavelengths. Important is the speed and exact reproducibility of the activation of a desired relative position of the focal plane 112.

FIG. 6 shows a multiple arrangement of lighting modules 100a-d, which form the illumination unit 100 in its entirety and their structured illumination pattern 20a-d or analog add their homogeneous illumination 21 to a larger illumination pattern 20 and a larger homogeneous illumination area. This should preferably be done without overlapping, since superposed illumination patterns 20a-d lead to contrast attenuation.

Each individual illumination module 100a-d has in each case a coherent light source 104a-d, an optional optic 106a-d for beam shaping when hitting a pattern generation element 108a-d and an imaging optic 110a-d. Since the power of a separate light source 104a-d is available to each subarea 20a-d, the total energy in the illumination pattern 20 is increased compared to a single identical light source 104. Thus, higher performance, ie in particular higher ranges can be achieved, or it comes with less expensive and simpler light sources lower output. Finally, the distribution of light output to multiple sources further simplifies any problems with eye protection limits.

For each individual module 100a-d, one of the previously individually described embodiments may be used, even in hybrid forms. It is also conceivable to use elements of the individual modules 100a-d multiple times, for example the pattern generating element 108. It can be accepted or prevented by appropriate arrangement and dimensioning that the interference of the multiple light sources 104a-d to the contrast impairing interfering superimpositions lead single pattern 20a-d.

All in all, a lighting concept will be presented, which allows a camera sensor based on stereoscopic images to capture the three-dimensional scenery safely and completely. The lighting concept provides very fast switching between a high contrast pattern structure providing projected image features for active stereoscopy and a uniform illumination intensity for utilizing natural scene contrast for detection. Under homogeneous lighting, a good visual orientation is possible, for which an additional camera would be needed in a purely active system with artificial structure.

A changeover would have to be realized conventionally by two different light sources, for example a contrast pattern illumination and a homogeneous illumination. You could also, about about a divider mirror, use the same imaging lens to select the lighting area. Nevertheless, a much higher cost of materials would be required.

Claims (12)

1. An illumination unit (100) having a light source (104) and having an optical pattern generating element (108) arranged in the beam path of the light source (104) for projecting an illumination pattern (20, 21) into a monitored zone (12),
   wherein an imaging objective (110) is disposed after the pattern generation element (108) and wherein an adjustment element (22, 110, 114, 116, 118) is provided by means of which the illumination pattern (20, 21) can be varied by setting the relative position of the focal plane (112) of the imaging objective (110),
   characterized in that
   the pattern generating element (108) has an optical phase element with irregularities which are designed to generate local phase differences between light portions incident in an adjacent manner onto the phase element and thus to generate the illumination pattern (20) by interference; and in that the setting element (22, 110, 114, 116, 118) is designed to set the relative position (112) of the focal plane of the imaging lens (110) so that the near field of the pattern generation element (108) is projected into the monitored zone (12) and generates an irregular pattern (20) and the structure of the illumination pattern (20, 21) can be varied by setting the relative position of the focal plane (112) of the imaging objective (110).
2. An illumination unit (100) in accordance with claim 1,
   wherein a switchover element (22) is provided to switch over between at least a first illumination (20) and a second illumination (21) of the monitored zone (12) different from the first illumination by varying the relative position of the focal plane (112) of the imaging objective (110).
3. An illumination unit (100) in accordance with claim 1 or claim 2, wherein the setting element (22, 110, 114, 116, 118) is designed to relocate the focal plane (112) into the pattern generation element (108) itself, in particular into the plane of a microstructure of the pattern generation element (108) and thus to generate a homogeneous intensity distribution (21) in the monitored zone (12).
4. An illumination unit (100) in accordance with any one of the preceding claims,
   wherein the setting element (22, 110, 114, 116, 118) has a focusing device of the imaging objective (110) and/or a spacing setting device (114, 116) between the imaging objective (110) and the pattern generation element (108), in particular for the relative displacement of the imaging objective (110) with respect to the pattern generation element (108) or for displacing a further optical element (118) between the pattern generation element (108) and the imaging objective (110).
5. An illumination unit (100) in accordance with any one of the preceding claims,
   wherein the light source (104) has at least one laser and wherein an optical distribution system which can expand the laser light for illuminating the pattern generation element (108) is provided between the laser and the pattern generating element (108) or an optical converging system (106) is provided to bundle the laser light onto the pattern generation element (108).
6. An arrangement of a plurality of illumination units (100a-d) in accordance with any one of the claims 1 to 5 for illuminating a larger monitored zone (12) or for generating a larger illumination intensity, wherein the illumination regions (20a-d) of the individual illumination units (100a-d) are added in an overlapping or nonoverlapping manner.
7. A stereoscopic 3D camera (10) having at least one illumination unit in accordance with any one of the claims 1 to 6, wherein an evaluation unit (22) is provided which is designed for the use of a stereo algorithm in which partial regions of the images associated with one another and taken by the two cameras (14a-b, 16a-b) of the stereoscopic camera (10) are recognized and their distance is calculated with reference to the disparity for generating a three-dimensional distance image.
8. A stereoscopic 3D camera (10) in accordance with claim 7,
   wherein the evaluation unit (22) is designed to generate a three-dimensional distance image beneath the first illumination (20) and the second illumination (21) and to compare the arising three-dimensional image data with one another or to offset them against one another.
9. A stereoscopic 3D camera (10) in accordance with claim 7 or claim 8,
   which is designed as a safety camera, wherein the evaluation unit (22) is able to recognize unauthorized intrusions into the monitored zone and can thereupon generate a switch-off signal, and wherein a safety output (26) is provided via which the safety camera can output a switch-off signal to a monitored machine.
10. A method for projecting an illumination pattern (20, 21) into a monitored zone (12) by means of a pattern generation element (108),
    wherein the illumination pattern (20) is generated by interference in that divergent coherent light is irradiated onto a phase element of the pattern generation element (108) and irregularities of the phase element (108) generate local phase differences between light portions incident in an adjacent manner onto the phase element and wherein the near field of the pattern generation element (108) is imaged in a first irregular illumination (20) into the monitored zone (12) by an imaging objective (110), an irregular pattern (20) is generated there and the structure of the illumination pattern (20, 21) is set by the relative position of the focal plane (112) of the imaging objective (110).
11. A method in accordance with claim 10,
    wherein a switchover is carried out between at least a first irregular illumination (20) and a second homogeneous illumination (21) in that the focal plane (112) of the imaging objective (110) is set for the first illumination (20) so that the near field of the pattern generation element (108) is projected into the monitored zone (12) and the focal plane (112) of the imaging objective (110) is set for the second illumination (21) onto the pattern generation element (108) itself and is in particular set into the plane of a microstructure of the pattern generation element (108).
12. A stereoscopic image detection method for the secure monitoring of a spatial zone (12) with the aid of a dense depth map of a scenery in a monitored zone (12) generated by a stereoscopic camera (10), wherein the monitored zone (12) is illuminated using a method in accordance with claim 11, wherein mutually associated part regions of the images taken by the two cameras (14a-b, 16a-b) of the stereoscopic camera (10) are recognized for generating the depth map and their spacing is calculated with reference to the disparity, and wherein a respective three-dimensional distance image is generated beneath the first illumination (20) and the second illumination (21) and the arising three-dimensional image data are compared with one another or offset against one another for generating the depth map, with the depth map being evaluated for unauthorized intrusions and, if an unauthorized intrusion is recognized, a switch-off signal being output to a danger source via a safety output (26).

Images